Although this seems like an enormous show of skill and awesome addition for the WebGPU ecosystem, as someone that hates anything AI represents or lightly touches, I do have to note that I am very much wary of the amount of AI and "vibe coding" used in the projects made by the author. The project's GitHub organization logo certainly is AI[3], tens of thousands of lines of code of full backend implementation introduced in a single commit[4] while the rest of the commit tree are a lot of weird small docs updates and version bumps. Not only the way the articles and readme's are written do feel like AI slop, the author also admitted to using AI to post entirely detailed responses[5] in another project's issues. Scratch that, those responses simply look like an AI chatbot posting with almost to no human interference.
I have no problem with people that need some level of assistance to better communicate outside of their native language, English is not my first language as well, but I've never needed to use AI for that purpose, and I'm sure people don't mind some mistakes coming from someone trying to learn the language. Hell, I will always prefer something written in broken English than some AI slop that reads with no soul. This project at first glance does seem incredibly impressive and exciting, but reeks of AI slop.
[0]: https://github.com/gogpu/naga [1]: https://github.com/gogpu/gg [2]: https://dev.to/kolkov/gogpu-a-pure-go-graphics-library-for-g... [3]: https://github.com/gogpu [4]: https://github.com/gogpu/wgpu/commit/6f0e34765777c41f94dfbb8... [5]: https://github.com/cogentcore/webgpu/issues/14#issuecomment-...
